Weifang Chen's research while affiliated with Zhejiang University and other places

... Following the previous research, 61 this study aims to further investigate the interaction between lateral jet flow with hypersonic rarefied crossflow. Using the newly developed NCCR model of multispecies thermochemical non-equilibrium gases, 59 the primary objective of this paper is to reveal the physical mechanism underlying lateral jet interaction with hypersonic rarefied crossflow, considering the real gas effect, and to explore the resulting impact on the performance of RCS in an in-flight vehicle. ...

... Subsequently, Myong simplified the convection term in the evolution equation of heat flux based on Eu's adiabatic hypothesis and closure theory [10], proposing a set of nonlinear coupled algebraic equations known as the nonlinear coupled constitutive relations (NCCR) model [11]. To date, the NCCR model has been extensively investigated and validated in various nonequilibrium gas flow problems [12][13][14][15][16][17]. The NCCR model can recover NS equation solutions in the continuum region and provide more accurate nonequilibrium phenomena and nonlinear effect predictions than traditional NS equations in several typical rarefied nonequilibrium flows, such as one-dimensional shock wave structure [14,18], Couette flow [19], Poiseuille gas flow [20], thermodynamic [21], and thermochemical nonequilibrium flows [12,22] along with threedimensional complex flows past hypersonic vehicles [23,24]. ...

... Regarding this type of multi-scale lateral jet flow, where the local thermodynamically equilibrium assumption fails, the NSF equations that are based on the continuum hypothesis may lose their validity to describe accurately the complex and strong non-equilibrium flowfield. 18 In addition, though the DSMC method is often regarded as the standard benchmark solution for the numerical simulations of hypersonic rarefied flows, the computational efficiency of the DSMC method is relatively low when simulating the near continuum flows. ...

... With the rapid advancement of aircraft and robotics technology 1 , the development of high-performance adaptive control 2 methods for uncertain systems in various scenarios has become a prominent research focus in recent decades 3,4,5 . Reviewing previous works 6,7,8,9,10,11 it is evident that the realization of a multi-task universal adaptive control system with autonomous learning capabilities poses significant challenges. ...

... These two methods calculate the gas distribution function at the cell interface through the local solution of the Boltzmann equation, which couples the free transport and the collision processes of gas molecules. The distribution function in UGKS is expressed by 28,29 f n ij;a ¼ e ÀDt=s f x ij À n a Dt; n a ; 0 ...

... Subsequently, Myong simplified the convection term in the evolution equation of heat flux based on Eu's adiabatic hypothesis and closure theory [10], proposing a set of nonlinear coupled algebraic equations known as the nonlinear coupled constitutive relations (NCCR) model [11]. To date, the NCCR model has been extensively investigated and validated in various nonequilibrium gas flow problems [12][13][14][15][16][17]. The NCCR model can recover NS equation solutions in the continuum region and provide more accurate nonequilibrium phenomena and nonlinear effect predictions than traditional NS equations in several typical rarefied nonequilibrium flows, such as one-dimensional shock wave structure [14,18], Couette flow [19], Poiseuille gas flow [20], thermodynamic [21], and thermochemical nonequilibrium flows [12,22] along with threedimensional complex flows past hypersonic vehicles [23,24]. ...

... Through these comparisons, it becomes evident that the SGH model yields better agreement with the DSMC data, particularly concerning the surface heat flux coefficient, when compared to the NS equations at Ma 10 and Kn 0.05. Article in Advance / TECHNICAL NOTES proximity to the wall, a recently established second-order slip/jump boundary condition modified by the SGH model [33] has been reported to extend the capability of the NS equations for moderately rarefied gas flows. ...

... Subsequently, Myong simplified the convection term in the evolution equation of heat flux based on Eu's adiabatic hypothesis and closure theory [10], proposing a set of nonlinear coupled algebraic equations known as the nonlinear coupled constitutive relations (NCCR) model [11]. To date, the NCCR model has been extensively investigated and validated in various nonequilibrium gas flow problems [12][13][14][15][16][17]. The NCCR model can recover NS equation solutions in the continuum region and provide more accurate nonequilibrium phenomena and nonlinear effect predictions than traditional NS equations in several typical rarefied nonequilibrium flows, such as one-dimensional shock wave structure [14,18], Couette flow [19], Poiseuille gas flow [20], thermodynamic [21], and thermochemical nonequilibrium flows [12,22] along with threedimensional complex flows past hypersonic vehicles [23,24]. ...

... It was shown that decreasing velocity component behavior was caused by the velocity slip parameter as it increased in both directions. Hui et al. (2023) investigated the velocity slip and temperature jump for a two-dimensional rough plate in hypersonic conditions using the Direct Simulation Monte Carlo (DSMC) technique. They came to the conclusion that altering the slip model increased the precision of macroscopic variables, particularly the coefficient of heat transmission, when compared to the conventional first-order slip boundary conditions. ...

... The first approach combines feature extraction algorithms with regression algorithms to perform a nonlinear mapping of local physical features for prediction. [3][4][5][6][7][8][9][10] Although it requires manual experts in extraction algorithms, such as determining the remained proper orthogonal decomposition (POD) modes by sorting eigenvalues or predefining the number of latent variables in variational autoencoder (VAE), 11 it is faster and requires fewer samples to train a model by this method. The second approach leverages the ability of deep learning algorithms to handle high-dimensional tensor data and constructs a mapping from initial tensor data to flow field tensor data. ...